Gi2MO Reader: A Windows 8 app with touch interface for Idea Management

Posted By on December 23, 2012

We are proud to announce another research prototype from Gi2MO Project: a Windows 8 application made for the new modern touch optimised UI from Microsoft (formally called Metro UI). The target devices for Gi2MO Reader are primary tablets and consuming content on the go.

Gi2MO Reader explores information presentation and interaction with Idea Management data delivered via the recently updated Gi2MO Ontology. Furthermore, similarly as our iPhone client, this application evaluates the use of IdeaStream Services – customisable REST API for accessing IdeaStream with RDF/XML as data exchange format.

The first basic version that we share today includes browsing multiple Gi2MO feeds and adding new feeds to the collection.

Main screen of Gi2MO Reader app with demo Gi2MO feeds loaded.

Furthermore, the app was developed as a training ground to play around with the new application design paradigm from Microsoft: writing desktop apps using JavaScript, HTML5 and CSS3. Some of the highlights development wise are:

  • loading data from local/remote static XML or REST services
  • asynchronous custom data sources that progressively load web content for ListView component(an example implementation in JavaScript for both item custom data source & custom group data source)
  • usage of CSS3 media tags to adjust application to snapped view
  • adding live tiles for each Gi2MO feed
  • basic use of application bar and application settings integrated with Windows 8 Modern UI
  • use of jQuery and RDFQuery for parsing RDF/XML in a Windows 8 app

Developing this app has also pushed and proven the flexibility of our IdeaStream Services solution based on RDFme data endpoint architecture. For the Gi2MO Reader we were required to create a REST service with pagination and customisable idea limit that returns the latest ideas plus all the statistics as with Gi2MO Stats app. We are happy to report that RDFme handled the challenge without any need for code update. For details see the SPARQL Plus code of the service and a sample RDF/XML generated with it.

Loading data from an individual Gi2MO feed using custom data source implementation and IdeaStream Services REST API.

For now, we did not put the Gi2MO Reader into the Windows Marketplace, however we share the source code together with the binaries so that it can be deployed on any Windows 8 PC (x86 or ARM). For details, more screenshots, as well as download links head over to the apps section.

Gi2MO Ontology v0.6: a close up look at idea relationships with Gi2MO Links hierarchy

Posted By on December 14, 2012

Gi2MO Ontology is a schema for Idea Management System metadata that we maintain as an implementation of the interchangeable data model proposed by the Gi2MO Project. Along the road we have shown a number of uses of the ontology for: distributed data visualisation, semantic search, data integration etc.

As a followup to our paper on idea similarity presented in October during the CollaborateCom conference, we introduce an update to the Gi2MO Ontology by adding an extended hierarchy of idea relationships: Gi2MO Links.

Establishing relationships in Idea Management System.

During our studies with various idea datasets like Ubuntu Brainstorm or Dell IdeaStorm we noticed that there is quite a lot more connections between ideas than just the currently collected duplicate information. Therefore, Gi2MO Links lists and names those missing relationship types, as well as puts them into a hierarchy that enables a variety of knowledge management opportunities.

Overview of the Gi2MO Links hierarchy for idea relationships.

In our case study, we analysed the usage of Gi2MO Links for idea clustering and information summarisation. The results were quite impressive giving us a 95% increase in the amount of detected relationships in comparison to duplicate detection. For more details see the paper, poster and presentation slides from the talk during CollaborateCom.

The hierarchy and a brief explanation of its usage can be seen under the modelling section. Additionally, all new relationships have been incorporated into the new version of the Gi2MO Ontology – Gi2MO v0.6.

Upcoming presentation on semantic search in Idea Management during ICITST 2012 in December

Posted By on November 13, 2012


Following on the two previous talks, we got one more event in the Gi2MO calendar for this year: a presentation on semantic search during the 7th International Conference for Internet Technology and Secured Transactions (ICITST-2012).

The talk is based on a paper by Geovanny Poveda, Adam Westerski and Carlos A. Iglesias entitled: “Application of Semantic Search in Idea Management Systems”. Here’s what you can expect:

  • overview of implementing a triple store as the backend of an Idea Management System
  • search over distributed Idea Management Systems using SPARQL query language
  • building user interfaces on top of SPARQL to add faceted browsing for Idea Management Systems

In more technical terms, this work involved a number of experiments with synchronising distributed IdeaStream instances using our RDFme solution and Gi2MO ontology to provide data integration. The experiments were based on integration of ARC2 triple store into Drupal and constructing some user interfaces on top of it (for example with Exhibit from SIMILE project for faceted browsing capabilities).

Lastly, we also did some experiments on scraping ideas from regular web content using a tool called Scrappy. The obtained data was structured according to Gi2MO Ontology and treated like a data source of any other Idea Management System (therefore enabling search together with normally submitted ideas via IdeaStream).

As a follow up of that research some time ago we actually released new RDFme version that allows sync between Idea Management Systems and distributed querying. If you follow the blog this was demoed with an iPhone app client.

RDFme module for IdeaStream: allows to construct Web Services based on SPARQL queries for ideas stored in the backend triple store.

All of those research and development experiences are going to be discussed during the talk. The conference is going to take place in December 10-12, 2012 in London, UK. The presentation slides and research paper will be uploaded to Gi2MO Docs afterwards so keep an eye out for updates!

Update: Conference final programme is out. Gi2MO will be presented during Internet Application and Technology session at 14:45, Monday, December 10th. Talk will be given by Geovanny Poveda.

Update2: The presentation slides and the accepted version of the paper are available now in Gi2MO Docs:

  • presentation slides [PDF]
  • paper accepted for the ICITST-2012 [PDF]

Gi2MO at ICT Proposers Day 2012

Posted By on September 24, 2012


The implementations and research of Gi2MO Project have contributed so far to a number of funded projects. Nevertheless, we are still engaging in new collaborations and starting new partnerships.

As part of that effort, we are going to be visting Warsaw from 26 to 27 September, during the ICT Proposers Day 2012 – an annual networking event that gathers researchers and entrepreneurs from Europe to collaborate on new research projects funded by the EU.

If you are participating in the event and interested in collaboration or just curious about our project be sure to contact us.

Below you can see Gi2MO leaflets that summarise some of the main contributions of the Gi2MO Project so far. You can also download the print ready version (cover and back for each leaflet) in a single PDF file.

Project Summary
Gi2MO Ontology
Opinion Mining
Gi2MO Types Taxonomy

Gi2MO going overseas to present at CollaborateCom conference!

Posted By on September 18, 2012

After quite some development oriented news we are happy to announce something more research flavoured. We are going to give a presentation about our recent research results at the 8th IEEE International Conference on Collaborative Computing: Networking, Applications and Worksharing (CollaborateCom 2012).

The talk is entitled “Idea Relationship Analysis in Open Innovation Crowdsourcing Systems” and will be given by Prof. Carlos A. Iglesias. Here is what you can expect:

  • idea relationship hierarchy that goes far beyond the currently used duplicate detection
  • insight into results of our annotation experiments and how new relationships perform in practice of Idea Management Systems
  • report on idea clustering capabilities of the new relationships
  • results of our study on connection between similarity of idea characteristics (e.g. innovation type) and the idea relationship type

The conference will be held in Pittsburgh, Pennsylvania, United States, starting on October 14th until 17th. The exact time and location of our talk will be posted when the schedule of the conference is available.

As usual, the slides and related research paper will be made available in our docs section at a later date. Be such to check back if you are interested in details of our findings!

Update: The conference has come to an end now, so we provide the presentation slides and the accepted version of the paper. Additionally, due to the courtesy of the organisers, we were able to present a demo and a poster as well:

  • presentation slides [PDF]
  • paper accepted for the CollaborateCom 2012 [PDF]
  • poster presented during the poster session in low resolution [PNG] or high resolution [PNG]
  • demo of IdeaStream Similarity software also presented during poster session [youtube]

Gi2MO poster presented during CollaborateCom 2012.

New version of IdeaStream (v1.6): Sandbox innovations before releasing them into the wild

Posted By on August 14, 2012


We are proud to bring you another update of our open-source Idea Management platform – Gi2MO IdeaSteam. This time we have new two major features and a bunch of improvements and fixes.

Firstly, we added a functionality called “Idea Sandbox”. What it does is to allow an inventor to save an unfinished idea without publishing it.

For example, if somebody wants to rethink his submission or add a better version before releasing it to the public, now he can store the idea in the system and come back to it at any time to edit it and publish when its ready. The sandbox is integrated with the Drupal account view, which now shows all published and unpublished ideas of the user.

New Idea Sandbox feature in IdeaStream v1.6.

The same functionality is available for Idea Contests and Idea Reviews. Furthermore, to give quick access to the Sandbox we placed a new link at the top of the screen (i.e. “My Ideas”).

The second feature that we added is support for Idea Reviews in the IdeaStream filter bar. Provided that the user has privileges to view the reviews he will see an additional option to browse the reviews and filter them by type. Additionally, we have improved the Idea Review snippet to make it more clear and in line with the design of Idea and Idea Contest snippets.

Idea Review browsing and filtering in IdeaStream v1.6.

If you would like to test the new IdeaStream v1.6 in action, you can download it directly from the IdeaStream website. For full release notes please refer to the install.txt file that you can find inside the download package.

If you are a user of IdeaStream Similarity or IdeaStream Analytics we recommend to update them to the newest versions in order to fully use the new features of Idea Sandbox.

Gi2MO Metrics: Making Sense of Taxonomy Annotations to Measure Idea Performance

Posted By on August 8, 2012

Some time ago we shared our research on Gi2MO Types – a taxonomy that allows to classify ideas based on a number of their characteristics. We also showed a sneak peak of how this could be used in practice. Today, we would like to go above and beyond that preview and deliver an actual solution that builds on top of our taxonomy – Gi2MO Metrics.

The principal notion behind Gi2MO metrics is to automatically generate a number of indicators that would allow to determine performance of idea contests, selected idea datasets or entire Idea Management Systems. The metrics are created based on the Gi2MO Types annotations (see here how to apply annotions in IdeaStream) – we propose one metric per each Gi2MO Types taxonomy sub-tree.

Gi2MO Metrics derived from annotations with Gi2MO Types taxonomy.

Apart of a theoretical solution, today we are also happy to deliver its implementation – Gi2MO Metrics support included in our IdeaStream Analytics module. The module is a supplementary component that calculates various statistics and gives access to data management panels for our open-source Idea Management System: Gi2MO IdeaStream. The support for Gi2MO Metrics is enabled automatically if you already installed Gi2MO Types taxonomy.

Gi2MO Metrics calculated by IdeaStream Analytics.

Before trying out the updated Analytics module, we recommend to check the detailed description of all metrics in a new subpage of Modelling section. If you wish to test how the metrics perform with your own data head over to the apps section and download the new release of IdeaStream Analytics module.

Apart of the Gi2MO Metrics on board, you may find out that the new IdeaStream Analytics version comes with a number of other updates: relationship management and statistics (for recently released IdeaStream Similarity module), CSV export for all available statistics and data management views.

Gi2MO IdeaStream Similarity v3.0 is out!

Posted By on August 2, 2012

Today we are happy to bring you an update of our similarity module for Gi2MO IdeaStream – an Open-Source Idea Management System based on Drupal. The version bump is quite significant (see last 2.0.2 release) but so are the new features!

Firstly, we have added similarity types: apart of the full-text search (word frequency), IdeaStream now allows to define similarity measures based on Drupal vocabularies. In other words, you can detect similar ideas based on their tags, idea categories or some custom taxonomies that you create.

IdeaStream Similarity: search types configuration screen (click to enlarge).

In addition, to choosing any Drupal vocabulary for similarity detection the administrator can set the allowed distance between terms in the taxonomy tree, and the threshold of how many terms of similar ideas should match. Furthermore, we allow mixing those search types, so for example you can set similar ideas to be detected only if they have: at least half tags matching, labelled under the same idea category and sort them by keyword frequency.

Second major feature, is directions for the relationship types. Now the administrator can mark relationships as inverse or symmetric. In practice, that means that now you can set a relationship that is an opposite of some other (for example “is duplicate of” , “has duplicate”). When similarity reports are submitted IdeaStream automatically takes care of all the logic behind and later allows you to analyse the relationships from any angle you wish.

IdeaStream Similarity: symmetric and inverse relationship types (click to enlarge).

Apart of those two mentioned features we have extended the number of configuration options, made lots of small enhancements of the user interface, and also applied some bug fixes. You can try it out yourself by downloading the module. For detailed explanations of new features please refer to the change log in the install.txt file as well as module manual in the IdeaStream Similarity dev page.

The module has changed quite significantly, so if you are migrating from old versions (e.g. v2.0.2) be sure to run the Drupal update.php script.

To make all those new exciting changes more digestible, we have put together a screencast that shows IdeaStream Similarity in action. All the credits for putting together the new version of the module as well as the screencast go to Javier Espinosa Garcia.

New webpage for IdeaStream

Posted By on July 15, 2012

The primary goal of Gi2MO project is research, however in parallel to our investigative work we have put quite some effort into development of software such as our open-source Idea Management System – Gi2MO IdeaStream.

We noticed that often people who come to us and want to use IdeaStream are just starting their adventure with innovation management. So often IdeaStream is being used as a tool to play around with ideation concepts and test how they work in practice. In those cases, for newcomers, the complexity of research solutions and information about our work on large scale installations can be overwhelming at the first glance.

For this reason, we are lunching an additional new website – exclusively for promoting Gi2MO IdeaStream!

New Gi2MO IdeaStream website.

The goal of the new IdeaStream homepage is to focus only on the software and explain the very basics of Idea Management as it is implemented by our platform. Keeping that in mind, the main Gi2MO website, as well as this blog, will still remain the central information point for our project where you will be able to find all updates on research as well as development.

Case study of extending Gi2MO Ontology for InnoWEB project

Posted By on June 4, 2012

Recently, we discussed extensions of the Gi2MO Ontology that fit it to the specifics of mobile devices as well as deliberation platforms in general. Now, we would like to share another story of an initiative independent from Gi2MO project but connecting with our ontology research.

InnoWEB is an ongoing project by the Mondragon University, ISEA institute and companies from Mondragon Corporation. One of the goals of the project is to establish and evaluate a collaborative innovation platform based on social software.

In this setting, the Gi2MO Ontology is being used for publishing idea metadata and sharing it with other web applications.

Mondragon researchers created a number of their own ontology extensions to map the specifics of innovation pipelines applied in the Mondragon cluster companies. The Mondragon Gi2MO Wave ontology reuses some of the existing Gi2MO concepts but also introduces a number of new ones. Most notably, a more strict view on Idea Descriptions by defining a particular set of allowed ones like Outcome, Target, Technology, Market etc.

The collaborative platform that Mondragon uses during their evaluations with companies is based on Drupal and as an extension of cooperation with Gi2MO project they also utilize our RDFme solution for publishing and consuming the data within the InnoWEB platform.

InnoWeb prototype with innovation waves interface (click to enlarge).

If you would like to know more about the results of Mondragon studdies and how they use Semantic Web metadata in InnoWEB, check out a related publication from the 17th International Conference on Concurrent Enterprising (ICE 2011): “A Case Study on the Use of Community Platforms for Inter-Enterprise Innovation” by Felix Larrinaga, Igor Santos, Osane Lizarralde, Alain Perez.

Additionally, you can see the specification of Gi2MO Wave extension published at our own website.